March 2008 Family Medicine
The relationship between moderate alcohol use and
health is complex. Over the last 30 years, a series of
observational studies have suggested a U- or L-shaped
relationship between alcohol use and coronary heart
disease (CHD), with the lowest risk among people
who drink approximately one drink daily.1, 2 In many of
these studies, all-cause mortality was also found to be
lowest among moderate drinkers.3-5 Moderate alcohol
use has also been associated with a lower incidence of
peripheral vascular disease,6, 7 cholelithiasis,8 and type
II diabetes mellitus.9
Several concerns have been raised about the studies
that link moderate alcohol use and CHD.10-14 Most im-
portantly, all of these studies have been observational,
raising the possibility that an unmeasured confounding
factor explains the association, and no long-term con-
trolled trial of the issue is apt to be completed in the
near future. Recent examples highlight the difficulty
that can occur in attempts to interpret the results of
In addition, even moderate alcohol consumption can
have risk. A pooled analysis of 322,647 women dem-
onstrated a dose-dependent increase in breast cancer
risk associated with moderate alcohol use.16 Alcohol use
also potentiates the hepatocellular injury of hepatitis
C, which could explain the increase in cirrhosis attrib-
utable to even moderate alcohol intake.4,17 Accidental
injuries are also associated with moderate drinking,18
and some investigators have found that moderate al-
cohol consumption may increase the risk of osteopo-
rotic fractures.19 Finally, concern has been raised that,
because of the habituating effect of alcohol, moderate
consumption could rise to an excessive level.20 At this
time, it is not known whether alcohol abuse is likely to
develop after a recommendation for increased alcohol
intake is made to a given individual.
Despite these controversies, we know little about
how patients view these concerns or whether they
Clinical Research and Methods
Beliefs, Motivations, and Opinions about Moderate
Drinking: A Cross-Sectional Survey
Kenneth J. Mukamal, MD, MPH, MA; Russell S. Phillips, MD; Murray A. Mittleman, MD
From the Division of General Medicine and Primary Care (Drs Mukamal
and Phillips) and the Division of Cardiology (Dr Mittleman), Beth Israel
Deaconess Medical Center, Boston, Mass.
Background: The relationship of moderate alcohol use and health remains controversial and un-
certain. How physicians and patients react to this uncertainty is unknown. Methods: We surveyed
outpatients at a single urban medical center that provides primary and tertiary care. Participants
completed a self-administered anonymous survey regarding their medical history, usual alcohol
consumption, and preferences and opinions regarding moderate drinking, defined as a drink every
1 to 2 days. All English-speaking individuals ages 21 years and older were eligible. Results: A total
of 878 outpatients participated, with a response rate of 79%. The median age was 47 years, and
57% were women. Approximately 60% of drinkers and 35% of abstainers agreed with the statement
that moderate drinking is a healthy activity and that it is safe for most people. About one third of
participants cited possible health benefits as part of their motivation for drinking alcohol. Those who
cited health benefits tended to be older, consumed alcohol more frequently but with a lower quantity
per drinking day, and were more likely to have a history of coronary heart disease. Only about 10%
of participants identified breast cancer as a possible risk of moderate drinking. When asked whether
they would be willing to consume one drink every 1–2 days if their doctor so recommended, 41%
of abstainers and 72% of all drinkers were willing to do so. Conclusions: A substantial number of
medical outpatients cite health benefits as a motivation for drinking alcohol and a willingness to
drink alcohol regularly if so recommended by a physician, although few recognize health risks from
(Fam Med 2008;40(3):188-95.)
189Vol. 40, No. 3
Clinical Research and Methods
would follow any proposed guideline that recommend-
ed moderate alcohol use. To explore patient preferences
and understanding regarding moderate alcohol, we
surveyed outpatients at a large urban medical center
that provides both primary and tertiary care to a diverse
sample of adults.
The Moderate Alcohol Use: Preferences and Out-
comes (MAHPO) survey queried a convenience sample
of outpatients at a large, private, nonprofit hospital in
Boston with approximately 530 licensed beds. The
hospital serves as a source of primary care for resi-
dents of eastern Massachusetts and as a tertiary care
center for patients throughout eastern Massachusetts
and surrounding regions. To maximize the clinical het-
erogeneity and geographic distribution of participants,
we included outpatients from multiple clinical units,
including three primary care practices and two car-
diology practices located at the hospital. The primary
care practices see more than 34,000 patients annually,
accounting for more than 70,000 visits. Of the 79% of
primary care patients who self-report their ethnicity,
66% are white, 20% are black, 5% are Hispanic, and
4% are Asian.
For this project, research assistants approached
patients in waiting areas and distributed and collected
surveys, which were completed entirely by participants.
Only one research assistant distributed surveys at any
one time. Surveys were anonymous, with no identifying
information. Exclusion criteria included age younger
than 21 years and inability to speak English. The sur-
vey was approved by the hospital’s institutional review
board with a waiver of written informed consent.
A signed introductory letter accompanied each sur-
vey. It described the survey as “for research about how
people feel about alcoholic beverages and how drinking
alcohol affects health.”
The survey assessed the following domains: demo-
graphics, medical history, current alcohol consumption,
beliefs regarding alcohol consumption, motivations
regarding drinking (ie, positive and negative features of
drinking alcohol), knowledge of medical consequences
of drinking, and willingness to change current alcohol
consumption. Medical history questions included items
about cardiovascular and liver disease and cardiovas-
cular risk factors. Family history of CHD was defined
as a myocardial infarction in a first-degree relative
before age 60.
Current alcohol intake was assessed in three ways.
First, we defined current drinkers by affirmative
responses to the question “Do you drink alcoholic
beverages?” with no specific time prompt. Second,
participants reported their usual quantity and frequency
of alcohol consumption, again with no specific time
prompt. Third, respondents completed the “high” ver-
sion of the TWEAK questionnaire (Tolerance, Worried,
Eye-openers, Amnesia, Kut down);21,22 the standard
TWEAK cutpoint for problem drinking is 3 points,
although 2 points has been used extensively.22,23
Beliefs regarding alcohol consumption and positive
and negative features of drinking alcohol were assessed
as a series of items in closed format without filtering.
Response options were formatted in Likert scales with
four categories ranging from strongly agree to strongly
disagree. The Cronbach α coefficients for these three
scales were 0.68, 0.77, and 0.85, respectively. Eleven
closed-format items regarding specific medical conse-
quences of having one drink every 1–2 days followed;
this level of intake was not labeled with any adjective,
such as “moderate.” Finally, willingness to change was
assessed with six items with yes/no response options
regarding personal willingness to drink regularly or
to abstain. The full instrument was pilot tested and
hand timed among otherwise eligible individuals prior
to formal administration to ensure its feasibility in its
We present categorical variables with counts and
frequencies, symmetric continuous variables with
means and standard deviations, and skewed continuous
variables with medians and interquartile ranges. We
performed univariate comparisons of binary variables
with Fisher exact tests, normally distributed continu-
ous variables with ANOVA, and skewed continuous
variables with Wilcoxon rank-sum scores.
Characteristics of Participants
Between 2002 and 2004, a total of 1,116 outpatients
were asked to participate in the survey. A total of 238
individuals declined to participate, for a response rate
of 79%. Characteristics of participants are shown in
Table 1. The median age was 47 years, with a range
from 21 to 90. Approximately 95% of respondents lived
in Massachusetts, their educational level was generally
high, and few were smokers. Cardiovascular risk factors
were relatively prevalent, but established cardiovascular
or liver disease was not. Sociodemographic charac-
teristics appeared to be representative of the medical
center’s primary care practice; 60.4% of the practice’s
population is female, and the median age is between
45 and 54 years.
Current Alcohol Intake
Approximately two thirds of respondents reported
current alcohol consumption (Table 1). Among current
drinkers, 267 (50%) drank less than weekly, 134 (25%)
drank 1–2 days per week, 90 (17%) drank 3–6 days
March 2008 Family Medicine
per week, and 45 (8%) drank daily. Median intake per
drinking day among current drinkers was two drinks
(interquartile range, 1–2.5). A total of 172 respondents
(20%), including similar proportions of drinkers and
abstainers, reported that an immediate family mem-
ber had been treated for alcoholism. TWEAK scores
ranged from 0 to 6, with 62 participants scoring 3 or
higher and another 118 scoring 2 points. A total of
29% of current drinkers with TWEAK scores of 3 or
higher reported that a physician had told them not to
drink any alcohol, compared with 5% of other current
Beliefs Regarding Alcohol Consumption
Table 2 shows the responses of participants to the
series of questions regarding alcohol consumption,
stratified by whether participants currently consumed
alcohol. Although abstainers and drinkers differed
significantly in their responses to all eight statements,
the patterns of response were most distinct for four
of the statements. As expected, 89% of the drinkers
agreed that they enjoyed drinking alcoholic beverages,
while 82% of the abstainers did not. Responses to
statements regarding whether moderate drinking is a
healthy activity and whether it is safe for most people
tended to track together (age- and sex-partial Spearman
r 0.54; P<.001); approximately 60% of drinkers and
35% of abstainers agreed with each statement, while
approximately 25% of abstainers strongly disagreed
with each. Finally, 18% of abstainers but only 5% of
drinkers strongly agreed that drinking alcohol in mod-
eration can lead to alcoholism. Men were significantly
more likely than women to agree with the “healthy”
and “safe” statements, but these differences were less
pronounced than noted for the comparison of drinkers
Motivations to Drink Alcohol
Table 3 shows the responses of current drinkers to
four statements regarding their motivations to drink
alcohol. Respondents most commonly cited enjoy-
ment of the taste of alcoholic beverages as a reason to
drink alcohol, although relaxation and its role in social
activities appeared nearly as common. About one third
of participants cited possible health benefits as part of
their motivation to drink alcohol.
We next compared the characteristics of the 188
drinkers who cited health benefits as a motivating
factor to the 334 who did not. Those who cited health
benefits tended to be older (mean age 49.9 ± 15.3 versus
42.6 ± 14.1 years, P<.001) and consumed alcohol more
frequently (2.3 ± 2.3 versus 1.6 ± 2.1 days per week,
P=.001) but with a lower quantity per drinking day
(median and interquartile range 1.5 (1.0–2.0) versus 2.0
(1.0–3.0) drinks per day, P<.001). They were somewhat
more likely to be male (51% versus 42%, P=.05) but
much more likely to have a history of coronary heart
disease (19% versus 9%, P=.003).
Table 4 presents respondents’ level of agreement to
specific motivations to limit their alcohol intake, strati-
fied by current consumption. In general, 50%–70%
of abstainers endorsed each statement with varying
levels of agreement; health problems, hangovers, and
driving were cited most often. Drinkers also com-
monly cited health problems, hangovers, and driving
as limiting factors in their alcohol intake, along with
sedation, but were less likely to cite such factors as
taste, gastrointestinal upset, medication interactions,
or fear of escalation.
Perceived Health Effects of Alcohol
Respondents reported whether they felt red wine,
white wine, beer, liquor, or no single beverage was
healthier than other alcoholic beverages. Among ab-
stainers, 92 (39%) endorsed no beverage as healthiest,
Characteristics of MAHPO Survey Respondents
Age in years
Less than high school
Some graduate school
Cardiovascular risk factors
Family history of premature MI
High blood pressure
MI or angina
CABG or PTCA
Any cardiovascular disease
Number of Respondents
MAHPO—Moderate Alcohol Use: Preferences and Outcomes survey
MI indicates myocardial infarction, CABG or PTCA indicates coronary
artery bypass surgery or angioplasty, and any cardiovascular disease
includes any of these and stroke. Current drinkers were defined by the
question “Do you drink alcoholic beverages?” with no specific time
191Vol. 40, No. 3
Clinical Research and Methods
Responses of Abstainers and Drinkers to Statements Regarding Their Beliefs About Alcohol
I would drink more if it were safe for me to do so.
I do not drink alcohol because of religious reasons.
I enjoy drinking alcoholic beverages.
Drinking alcohol in moderation can lead to alcoholism.
I believe that drinking is safe for most people.
Drinking alcohol in moderation is a healthy activity.
Drinking alcohol, even in moderation, is a sign of weakness
in a person.
Drinking any alcohol is a sin.
* Total numbers vary due to nonresponse to selected questions.
** P values derive from exact tests
Responses of Drinkers to Statements Regarding Their Motivations to Drink Alcohol
I drink alcohol in part because . . .
I enjoy the taste.
It helps me to relax.
Drinking is part of having fun with friends and family.
It may prevent health problems.
* Total numbers vary due to nonresponse
March 2008 Family Medicine
130 (56%) endorsed red wine, and 11 (5%) endorsed
another option. In contrast, among drinkers, 102 (20%)
endorsed no single beverage, 392 (76%) endorsed red
wine, and 21 (4%) endorsed another option (P<.001).
Men and women did not differ in their responses
Respondents were asked whether they felt consumers
of one drink every 1–2 days had shorter or longer lives
than abstainers. A total of 43% of abstainers and 62% of
current drinkers indicated that moderate drinkers lived
longer lives (P<.001); the corresponding proportions
of men and women were 63% and 51%, respectively
Table 5 describes respondents’ beliefs regarding the
associations of moderate drinking with specific health
conditions. Only a minority of participants reported that
they perceived that moderate drinking prevented any
health condition, even myocardial infarction. Slightly
more than half of both abstainers and drinkers believed
that one drink every 1–2 days could cause liver dam-
age and birth defects. Abstainers were approximately
twice as likely to believe that that level of intake caused
myocardial infarction, stroke, gallstones, and diabetes,
although the proportions varied somewhat. Only about
10% of all participants identified breast cancer as a pos-
sible risk of moderate drinking; among women, 14% of
abstainers and 11% of drinkers identified this risk.
Willingness to Change Alcohol Intake
Lastly, respondents reported their willingness to
change alcohol consumption under various scenarios.
Among 127 individuals who reported alcohol con-
Responses of Abstainers and Drinkers to Statements
Regarding Their Motivations to Limit Their Alcohol Intake
I don’t drink more alcohol than I do because . . .
I dislike the taste.
It gives me an upset stomach or heartburn.
It costs too much.
It makes me lose concentration or become sleepy.
It interacts with medicines that I take.
It affects my ability to drive.
It causes health problems.
I fear losing control of my drinking habit.
It causes “hangovers.”
* Total numbers vary due to nonresponse